1. Field
Example embodiments of the present invention relate in general to terrain representation, and more specifically, to a method for representing a terrain based on terrain primitives, a method for creating terrain primitives, and an apparatus using the methods.
2. Description of Related Art
Lately, creation of a large-scale realistic terrain is required in various fields of content, such as movies, games, and the like. In order to obtain terrain data for creating a large-scale realistic terrain, various methods have been used.
A simplest method is to enable a user to generate a terrain by himself/herself by controlling the height of a terrain using a terrain brush. However, the method is costly and takes a long time since a user should generate a terrain manually. In order to generate a realistic, high-quality terrain while minimizing a time and cost for generating a terrain, methods of automatically generating a terrain using a computer have been proposed.
One of the proposed methods is procedural terrain generation. The procedural terrain generation is a technique of repeatedly applying a predetermined rule or method to generate a complex terrain. As examples of the procedural terrain generation, there are a Perlin noise method, a fractal noise method, etc.
The Perlin noise method, which is a representative noise-based terrain modeling method, involves generating a terrain using Perlin noise. In detail, the Perlin noise method includes generating random numbers through a noise function, and applying the random numbers to a height map to thereby generate a terrain. The random numbers can generate new wavelengths by amplitudes and frequencies so that various shapes of terrains can be represented by a combination of wavelengths.
Meanwhile, the fractal noise method involves generating a terrain using fractal noise. A representative terrain modeling method using fractal noise is a midpoint displacement method, and the midpoint displacement method is based on a diamond-square algorithm. According to the diamond-square algorithm, a center point is calculated using four corner points, and a middle point is calculated using the center point and the corner points.
The procedural terrain generation can easily generate a terrain using a minimum rule, but has difficulties in generating a user's desired shape of a terrain.
In order to overcome the disadvantage of the procedural terrain generation, a sketch-based terrain generation method of building a user's desired land-form through a rough sketch by the user, and enhancing regional reality of a terrain through a procedural method and an erosion/weathering simulation technique has been proposed. The sketch-based terrain generation method helps build a land-form through sketch lines but has limitations in representing a regional concave-convex structure.
As an automated terrain generation method other than the procedural terrain generation, there is a technique based on a digital elevation model (DEM). The DEM is image data extracted from contours, aerophotos, satellite images, and the like, and represents a height of a given area. The DEM is realistic since it is built by extracting a real terrain. However, it is nearly impossible to obtain a DEM that exactly represents an imaginary terrain intended by a user.
In order to overcome the disadvantages of the terrain generation methods as described above, a combination of the terrain generation methods has been proposed. For example, there is a method of easily creating a realistic shape of a terrain intended by a user by extracting features of ridges or valleys from a DEM, splitting a terrain into terrain patches based on the features to rearrange the terrain patches according to ridges or gorges intended by the user, and smoothly synthesizing the terrain patches.
However, this method cannot ensure realistic representation of the automatically synthesized terrain either.